[address-policy-wg] Input Requested: How to Ensure Responsible ASN Resource Management

Dear Elvis, dear colleagues,

We are happy to provide you with the data you requested.

In the last 12 months, about 18% (414 out of 2291) of the requests were 
for 16-bit ASNs. Currently we have enough 16-bit ASNs in our free pool, 
but it should be noted that the return of unused ASNs is the only source 
to add such ASNs to our pool.

Once an ASN has been returned to RIPE NCC, the only reason to keep it in 
our quarantine pool for longer than 6 months is that the ASN is still 
visible in the routing system. We currently have just over 200 16-bit 
ASNs in regular quarantine, mainly thanks to our project to clean up 
unused ASNs.

It is correct that we still assign 16-bit ASNs when they are 
specifically requested. Although the ASN policy states that since "2010 
the RIPE NCC will cease to make any distinction", it was decided in 2010 
to keep the option to request for 16-bit ASNs [1][2]
Neither IANA nor other RIRs make this distinction between 16-bit and 
32-bit ASNs any more.

I hope you found this information useful, especially in terms of how to 
achieve more responsible ASN management.

Kind regards,
Marco Schmidt
Manager Registration Services
RIPE NCC

[1] https://www.ripe.net/publications/docs/ripe-679#ASnumbers
[2] 
https://www.ripe.net/ripe/mail/archives/address-policy-wg/2010-January/004977.html

On 12/06/2023 12:46, Elvis Daniel Velea wrote:
> Hi Nick,
>
> On Mon, Jun 12, 2023 at 00:25 Nick Hilliard <nick at foobar.org> wrote:
>
>     Elvis Daniel Velea wrote on 11/06/2023 10:06:
>     > If someone could tell me why a 32bit ASN can not be used today
>     (even
>     > with 10 years old equipment), I’d appreciate it.
>
>     there's plenty of kit out there which still doesn't support BGP Large
>     Communities, particularly mikrotik routeros which only released an
>     initial production implementation at around the beginning of 2022.
>
>     Nick
>
>
> thanks for that. I’m surprised to hear this, but heh, manufacturers 
> can be slow sometimes… very slow :(
>
> It may, then, matter to some if an ASN is 16bit or not. I know that 
> the NCC had at some point (maybe still do) assigned 16bit ASNs upon 
> request. Curious to see some stats, if possible:
> - how many requests come in for 16bit a year? how many are those of 
> the total ASN requests?
> - how many 16bit ASNs are still in the pool and how many come back to 
> the free pool every year?
>
> Just trying to see how many years until 16bit ASNs could only be 
> issued by the NCC only if returned.
>
> On another note, I believe that there were a lot of 16bit ASNs in 
> quarantine because of references in various objects (mostly as-sets if 
> I recall correctly). Can you tell us, Marco, how many of those ASNs 
> are quarantined and why aren’t these removed out of quarantine and 
> assigned?
>
> elvis
>
> -- 
> This message was sent from a mobile device. Some typos may be possible.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: </ripe/mail/archives/address-policy-wg/attachments/20230612/5f4f61a0/attachment.html>